Transcript Psychology 3260: Personality & Social Development

Psychology 3260:
Personality & Social
Development
Don Hartmann
Spring 2006
Supplementary Lecture 2: © Method I
1
Supplementary Readings


Hartmann, D. P., & Pelzel, K. (in press). Design,
measurement, and analysis in developmental
research. In M. H. Bornstein & M. Lamb (Eds.),
Developmental psychology: An advanced
textbook (5th. ed., pp.). Mahwah, NJ: Lawrence
Erlbaum.
Miller, S. A. (1998). Developmental research
methods (2nd ed.). Englewoods Cliffs, NJ:
Prentice-Hall.
2
Overview
Methods of Study. The Context:
Should we let kids watch violence on
TV (pp. 388-396)?
 Preparing for class discussion of a
current event—if time allows

3
Context: Violence in the Media
Violence in the media is an important social
problem with extensive history
Ancient Greeks
Plato versus Aristotle
Contemporary Theorists:
Freud, Lorenz: the catharsis hypothesis
Bandura et al.
Which theory is right? Let the data speak…
4
Theories & Facts
5
The Case Study


More-or-less careful descriptions of the
natural experiences of a single individual.
E.g., case study described in the newspaper
of a child who observed a movie in which the
protagonist fed ground glass to the victim by
disguising it in candy; the child viewer
subsequently attempted to do the same thing
to a dislike teacher.
6
Evaluating the Case Study
Value: Suggests possible relationships
(Heuristic value)

Dangers:
Accuracy of the description (bias or
forgetfulness) on the part of the reporter
Lying on the part of the individual
The question of causality: What causes
what?
Concerns about generality: Just who
would are the results generalizable to?
Note: Dangers better known as Validity
Threats—after Campbell et al.
7
Correlational
Investigations


r
Better name is ex post facto studies or
studies of natural variation
Investigator measures the two variables to
be correlated (e.g., viewing TV violence &
aggression) in a sample and correlates the
two variables (e.g., study by Eron et al.)
8
The Eron, Huesmann et al.
Study




Studied 875 8th graders in rural upstate
NY
Parental reports of TV viewing habits (a
retrospective report)
Peer nominations used to measure the
children’s aggressiveness
Longitudinal design (psychologists)
better called cross-lagged panel design
(sociologists)
9
r
r
More on Correlations


Measure of strength of relationship or association
with range from -1.0 to +1.0
r
Interpretations:
 -1.0 (perfect negative relationship): High score on
one variable invariably goes with low score on
the other (e.g., distance of nose from ceiling and
nose from floor; % right and % wrong).
r
 Other negative correlations: High scores on one
variable tend to go with low scores on the other
(e.g., golf and bowling scores). r
r
10
r
r
r
Still more on Correlations
r

0.0: No relationship (e.g., nose length and IQ)
r
r
r
Other positive correlations: High score on one
variable tend to go with high score on the other
+1.0 (e.g., height and IQ).
r
r
+1.0 (perfect positive relationship): High
score on one variable invariably go with high
score on the other (e.g., height in inches and
in centimeters).
r
r
r
r
r
11
And Correlations are Statistics
"The men are excited about shooting a statistician."
statatician."
12
Evaluating Ex post facto
Designs





Correlation is a necessary--but not sufficient-condition for causation
Possibilities if X (e.g., viewing violent TV) & Y (acting
aggressively) are correlated ( means causes):
XY
YX
ZX&Y
13
Summary of: Methods I


First shot at Methods
Go in Peace
14